forked from dbchaincloud/tendermint
-
Notifications
You must be signed in to change notification settings - Fork 0
/
pool.go
596 lines (509 loc) · 14.2 KB
/
pool.go
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
package blockchain
import (
"errors"
"fmt"
"math"
"sync"
"sync/atomic"
"time"
cmn "github.com/tendermint/tendermint/libs/common"
flow "github.com/tendermint/tendermint/libs/flowrate"
"github.com/tendermint/tendermint/libs/log"
"github.com/tendermint/tendermint/p2p"
"github.com/tendermint/tendermint/types"
)
/*
eg, L = latency = 0.1s
P = num peers = 10
FN = num full nodes
BS = 1kB block size
CB = 1 Mbit/s = 128 kB/s
CB/P = 12.8 kB
B/S = CB/P/BS = 12.8 blocks/s
12.8 * 0.1 = 1.28 blocks on conn
*/
const (
requestIntervalMS = 2
maxTotalRequesters = 600
maxPendingRequests = maxTotalRequesters
maxPendingRequestsPerPeer = 20
// Minimum recv rate to ensure we're receiving blocks from a peer fast
// enough. If a peer is not sending us data at at least that rate, we
// consider them to have timedout and we disconnect.
//
// Assuming a DSL connection (not a good choice) 128 Kbps (upload) ~ 15 KB/s,
// sending data across atlantic ~ 7.5 KB/s.
minRecvRate = 7680
// Maximum difference between current and new block's height.
maxDiffBetweenCurrentAndReceivedBlockHeight = 100
)
var peerTimeout = 15 * time.Second // not const so we can override with tests
/*
Peers self report their heights when we join the block pool.
Starting from our latest pool.height, we request blocks
in sequence from peers that reported higher heights than ours.
Every so often we ask peers what height they're on so we can keep going.
Requests are continuously made for blocks of higher heights until
the limit is reached. If most of the requests have no available peers, and we
are not at peer limits, we can probably switch to consensus reactor
*/
type BlockPool struct {
cmn.BaseService
startTime time.Time
mtx sync.Mutex
// block requests
requesters map[int64]*bpRequester
height int64 // the lowest key in requesters.
// peers
peers map[p2p.ID]*bpPeer
maxPeerHeight int64
// atomic
numPending int32 // number of requests pending assignment or block response
requestsCh chan<- BlockRequest
errorsCh chan<- peerError
}
func NewBlockPool(start int64, requestsCh chan<- BlockRequest, errorsCh chan<- peerError) *BlockPool {
bp := &BlockPool{
peers: make(map[p2p.ID]*bpPeer),
requesters: make(map[int64]*bpRequester),
height: start,
numPending: 0,
requestsCh: requestsCh,
errorsCh: errorsCh,
}
bp.BaseService = *cmn.NewBaseService(nil, "BlockPool", bp)
return bp
}
func (pool *BlockPool) OnStart() error {
go pool.makeRequestersRoutine()
pool.startTime = time.Now()
return nil
}
func (pool *BlockPool) OnStop() {}
// Run spawns requesters as needed.
func (pool *BlockPool) makeRequestersRoutine() {
for {
if !pool.IsRunning() {
break
}
_, numPending, lenRequesters := pool.GetStatus()
if numPending >= maxPendingRequests {
// sleep for a bit.
time.Sleep(requestIntervalMS * time.Millisecond)
// check for timed out peers
pool.removeTimedoutPeers()
} else if lenRequesters >= maxTotalRequesters {
// sleep for a bit.
time.Sleep(requestIntervalMS * time.Millisecond)
// check for timed out peers
pool.removeTimedoutPeers()
} else {
// request for more blocks.
pool.makeNextRequester()
}
}
}
func (pool *BlockPool) removeTimedoutPeers() {
pool.mtx.Lock()
defer pool.mtx.Unlock()
for _, peer := range pool.peers {
if !peer.didTimeout && peer.numPending > 0 {
curRate := peer.recvMonitor.Status().CurRate
// curRate can be 0 on start
if curRate != 0 && curRate < minRecvRate {
err := errors.New("peer is not sending us data fast enough")
pool.sendError(err, peer.id)
pool.Logger.Error("SendTimeout", "peer", peer.id,
"reason", err,
"curRate", fmt.Sprintf("%d KB/s", curRate/1024),
"minRate", fmt.Sprintf("%d KB/s", minRecvRate/1024))
peer.didTimeout = true
}
}
if peer.didTimeout {
pool.removePeer(peer.id)
}
}
}
func (pool *BlockPool) GetStatus() (height int64, numPending int32, lenRequesters int) {
pool.mtx.Lock()
defer pool.mtx.Unlock()
return pool.height, atomic.LoadInt32(&pool.numPending), len(pool.requesters)
}
// TODO: relax conditions, prevent abuse.
func (pool *BlockPool) IsCaughtUp() bool {
pool.mtx.Lock()
defer pool.mtx.Unlock()
// Need at least 1 peer to be considered caught up.
if len(pool.peers) == 0 {
pool.Logger.Debug("Blockpool has no peers")
return false
}
// Some conditions to determine if we're caught up.
// Ensures we've either received a block or waited some amount of time,
// and that we're synced to the highest known height. Note we use maxPeerHeight - 1
// because to sync block H requires block H+1 to verify the LastCommit.
receivedBlockOrTimedOut := pool.height > 0 || time.Since(pool.startTime) > 5*time.Second
ourChainIsLongestAmongPeers := pool.maxPeerHeight == 0 || pool.height >= (pool.maxPeerHeight-1)
isCaughtUp := receivedBlockOrTimedOut && ourChainIsLongestAmongPeers
return isCaughtUp
}
// We need to see the second block's Commit to validate the first block.
// So we peek two blocks at a time.
// The caller will verify the commit.
func (pool *BlockPool) PeekTwoBlocks() (first *types.Block, second *types.Block) {
pool.mtx.Lock()
defer pool.mtx.Unlock()
if r := pool.requesters[pool.height]; r != nil {
first = r.getBlock()
}
if r := pool.requesters[pool.height+1]; r != nil {
second = r.getBlock()
}
return
}
// Pop the first block at pool.height
// It must have been validated by 'second'.Commit from PeekTwoBlocks().
func (pool *BlockPool) PopRequest() {
pool.mtx.Lock()
defer pool.mtx.Unlock()
if r := pool.requesters[pool.height]; r != nil {
/* The block can disappear at any time, due to removePeer().
if r := pool.requesters[pool.height]; r == nil || r.block == nil {
PanicSanity("PopRequest() requires a valid block")
}
*/
r.Stop()
delete(pool.requesters, pool.height)
pool.height++
} else {
panic(fmt.Sprintf("Expected requester to pop, got nothing at height %v", pool.height))
}
}
// Invalidates the block at pool.height,
// Remove the peer and redo request from others.
// Returns the ID of the removed peer.
func (pool *BlockPool) RedoRequest(height int64) p2p.ID {
pool.mtx.Lock()
defer pool.mtx.Unlock()
request := pool.requesters[height]
peerID := request.getPeerID()
if peerID != p2p.ID("") {
// RemovePeer will redo all requesters associated with this peer.
pool.removePeer(peerID)
}
return peerID
}
// TODO: ensure that blocks come in order for each peer.
func (pool *BlockPool) AddBlock(peerID p2p.ID, block *types.Block, blockSize int) {
pool.mtx.Lock()
defer pool.mtx.Unlock()
requester := pool.requesters[block.Height]
if requester == nil {
pool.Logger.Info("peer sent us a block we didn't expect", "peer", peerID, "curHeight", pool.height, "blockHeight", block.Height)
diff := pool.height - block.Height
if diff < 0 {
diff *= -1
}
if diff > maxDiffBetweenCurrentAndReceivedBlockHeight {
pool.sendError(errors.New("peer sent us a block we didn't expect with a height too far ahead/behind"), peerID)
}
return
}
if requester.setBlock(block, peerID) {
atomic.AddInt32(&pool.numPending, -1)
peer := pool.peers[peerID]
if peer != nil {
peer.decrPending(blockSize)
}
} else {
pool.Logger.Info("invalid peer", "peer", peerID, "blockHeight", block.Height)
pool.sendError(errors.New("invalid peer"), peerID)
}
}
// MaxPeerHeight returns the highest height reported by a peer.
func (pool *BlockPool) MaxPeerHeight() int64 {
pool.mtx.Lock()
defer pool.mtx.Unlock()
return pool.maxPeerHeight
}
// Sets the peer's alleged blockchain height.
func (pool *BlockPool) SetPeerHeight(peerID p2p.ID, height int64) {
pool.mtx.Lock()
defer pool.mtx.Unlock()
peer := pool.peers[peerID]
if peer != nil {
peer.height = height
} else {
peer = newBPPeer(pool, peerID, height)
peer.setLogger(pool.Logger.With("peer", peerID))
pool.peers[peerID] = peer
}
if height > pool.maxPeerHeight {
pool.maxPeerHeight = height
}
}
func (pool *BlockPool) RemovePeer(peerID p2p.ID) {
pool.mtx.Lock()
defer pool.mtx.Unlock()
pool.removePeer(peerID)
}
func (pool *BlockPool) removePeer(peerID p2p.ID) {
for _, requester := range pool.requesters {
if requester.getPeerID() == peerID {
requester.redo(peerID)
}
}
delete(pool.peers, peerID)
}
// Pick an available peer with at least the given minHeight.
// If no peers are available, returns nil.
func (pool *BlockPool) pickIncrAvailablePeer(minHeight int64) *bpPeer {
pool.mtx.Lock()
defer pool.mtx.Unlock()
for _, peer := range pool.peers {
if peer.didTimeout {
pool.removePeer(peer.id)
continue
}
if peer.numPending >= maxPendingRequestsPerPeer {
continue
}
if peer.height < minHeight {
continue
}
peer.incrPending()
return peer
}
return nil
}
func (pool *BlockPool) makeNextRequester() {
pool.mtx.Lock()
defer pool.mtx.Unlock()
nextHeight := pool.height + pool.requestersLen()
if nextHeight > pool.maxPeerHeight {
return
}
request := newBPRequester(pool, nextHeight)
pool.requesters[nextHeight] = request
atomic.AddInt32(&pool.numPending, 1)
err := request.Start()
if err != nil {
request.Logger.Error("Error starting request", "err", err)
}
}
func (pool *BlockPool) requestersLen() int64 {
return int64(len(pool.requesters))
}
func (pool *BlockPool) sendRequest(height int64, peerID p2p.ID) {
if !pool.IsRunning() {
return
}
pool.requestsCh <- BlockRequest{height, peerID}
}
func (pool *BlockPool) sendError(err error, peerID p2p.ID) {
if !pool.IsRunning() {
return
}
pool.errorsCh <- peerError{err, peerID}
}
// unused by tendermint; left for debugging purposes
func (pool *BlockPool) debug() string {
pool.mtx.Lock()
defer pool.mtx.Unlock()
str := ""
nextHeight := pool.height + pool.requestersLen()
for h := pool.height; h < nextHeight; h++ {
if pool.requesters[h] == nil {
str += fmt.Sprintf("H(%v):X ", h)
} else {
str += fmt.Sprintf("H(%v):", h)
str += fmt.Sprintf("B?(%v) ", pool.requesters[h].block != nil)
}
}
return str
}
//-------------------------------------
type bpPeer struct {
pool *BlockPool
id p2p.ID
recvMonitor *flow.Monitor
height int64
numPending int32
timeout *time.Timer
didTimeout bool
logger log.Logger
}
func newBPPeer(pool *BlockPool, peerID p2p.ID, height int64) *bpPeer {
peer := &bpPeer{
pool: pool,
id: peerID,
height: height,
numPending: 0,
logger: log.NewNopLogger(),
}
return peer
}
func (peer *bpPeer) setLogger(l log.Logger) {
peer.logger = l
}
func (peer *bpPeer) resetMonitor() {
peer.recvMonitor = flow.New(time.Second, time.Second*40)
initialValue := float64(minRecvRate) * math.E
peer.recvMonitor.SetREMA(initialValue)
}
func (peer *bpPeer) resetTimeout() {
if peer.timeout == nil {
peer.timeout = time.AfterFunc(peerTimeout, peer.onTimeout)
} else {
peer.timeout.Reset(peerTimeout)
}
}
func (peer *bpPeer) incrPending() {
if peer.numPending == 0 {
peer.resetMonitor()
peer.resetTimeout()
}
peer.numPending++
}
func (peer *bpPeer) decrPending(recvSize int) {
peer.numPending--
if peer.numPending == 0 {
peer.timeout.Stop()
} else {
peer.recvMonitor.Update(recvSize)
peer.resetTimeout()
}
}
func (peer *bpPeer) onTimeout() {
peer.pool.mtx.Lock()
defer peer.pool.mtx.Unlock()
err := errors.New("peer did not send us anything")
peer.pool.sendError(err, peer.id)
peer.logger.Error("SendTimeout", "reason", err, "timeout", peerTimeout)
peer.didTimeout = true
}
//-------------------------------------
type bpRequester struct {
cmn.BaseService
pool *BlockPool
height int64
gotBlockCh chan struct{}
redoCh chan p2p.ID //redo may send multitime, add peerId to identify repeat
mtx sync.Mutex
peerID p2p.ID
block *types.Block
}
func newBPRequester(pool *BlockPool, height int64) *bpRequester {
bpr := &bpRequester{
pool: pool,
height: height,
gotBlockCh: make(chan struct{}, 1),
redoCh: make(chan p2p.ID, 1),
peerID: "",
block: nil,
}
bpr.BaseService = *cmn.NewBaseService(nil, "bpRequester", bpr)
return bpr
}
func (bpr *bpRequester) OnStart() error {
go bpr.requestRoutine()
return nil
}
// Returns true if the peer matches and block doesn't already exist.
func (bpr *bpRequester) setBlock(block *types.Block, peerID p2p.ID) bool {
bpr.mtx.Lock()
if bpr.block != nil || bpr.peerID != peerID {
bpr.mtx.Unlock()
return false
}
bpr.block = block
bpr.mtx.Unlock()
select {
case bpr.gotBlockCh <- struct{}{}:
default:
}
return true
}
func (bpr *bpRequester) getBlock() *types.Block {
bpr.mtx.Lock()
defer bpr.mtx.Unlock()
return bpr.block
}
func (bpr *bpRequester) getPeerID() p2p.ID {
bpr.mtx.Lock()
defer bpr.mtx.Unlock()
return bpr.peerID
}
// This is called from the requestRoutine, upon redo().
func (bpr *bpRequester) reset() {
bpr.mtx.Lock()
defer bpr.mtx.Unlock()
if bpr.block != nil {
atomic.AddInt32(&bpr.pool.numPending, 1)
}
bpr.peerID = ""
bpr.block = nil
}
// Tells bpRequester to pick another peer and try again.
// NOTE: Nonblocking, and does nothing if another redo
// was already requested.
func (bpr *bpRequester) redo(peerId p2p.ID) {
select {
case bpr.redoCh <- peerId:
default:
}
}
// Responsible for making more requests as necessary
// Returns only when a block is found (e.g. AddBlock() is called)
func (bpr *bpRequester) requestRoutine() {
OUTER_LOOP:
for {
// Pick a peer to send request to.
var peer *bpPeer
PICK_PEER_LOOP:
for {
if !bpr.IsRunning() || !bpr.pool.IsRunning() {
return
}
peer = bpr.pool.pickIncrAvailablePeer(bpr.height)
if peer == nil {
//log.Info("No peers available", "height", height)
time.Sleep(requestIntervalMS * time.Millisecond)
continue PICK_PEER_LOOP
}
break PICK_PEER_LOOP
}
bpr.mtx.Lock()
bpr.peerID = peer.id
bpr.mtx.Unlock()
// Send request and wait.
bpr.pool.sendRequest(bpr.height, peer.id)
WAIT_LOOP:
for {
select {
case <-bpr.pool.Quit():
bpr.Stop()
return
case <-bpr.Quit():
return
case peerID := <-bpr.redoCh:
if peerID == bpr.peerID {
bpr.reset()
continue OUTER_LOOP
} else {
continue WAIT_LOOP
}
case <-bpr.gotBlockCh:
// We got a block!
// Continue the for-loop and wait til Quit.
continue WAIT_LOOP
}
}
}
}
//-------------------------------------
type BlockRequest struct {
Height int64
PeerID p2p.ID
}